Showing posts with label sobriety (at 5:00 PM). Show all posts
Showing posts with label sobriety (at 5:00 PM). Show all posts

Wednesday, November 9, 2016

Actually, the Pollsters Got it Largely Right

Miracle man

Do not misunderstand what you are about to read. I am ecstatic that Donald Trump won. I probably woke the whole condo building up with my 2:30 AM "woo-hoo!"s. A Hillary Clinton victory would have been a disaster for the good people of this country (and the good people of many other countries). Conversely, I think Trump will make a fine president. And I believe he's our best shot at curing us from eight years of Obama as well as the political correctness now pressing down on us like a boot.

However, some sober analysis is in order. (Like I should talk. I woke up this morning with one of the biggest hangovers of my adult life.)

One of the headlines on Drudge claims:
Pollsters suffer biggest embarrassment in history
It's true that the Trump victory embarrassed many. It also surprised many (including me).

But the polls in the final days were actually pretty accurate.

On the eve of the election, the Real Clear Politics poll averages put Clinton ahead in the popular vote by 2.6%. She ended up winning the popular vote (probably - it was so close that this may not be confirmed for weeks) by perhaps .1%. Coming within 2.5% isn't bad for a poll, or an average of polls. It certainly isn't embarrassing.

Ironically, the polls that had Clinton winning the popular vote were actually closer to the actual margin than the most famous poll to have predicted the Trump victory. (The "pro-Clinton" polls were 3.6% off - the LA Times/USC Tracking poll was 5% off).

Two days ago, I posted a Real Clear Politics electoral map based on who was favored in the state polls. The map put Clinton ahead by 6 delegates, meaning that Trump had to "flip" at least one state.


He ended up flipping three (probably - Michigan might still flip back), and Clinton might have flipped one herself (Arizona, which still hasn't been called). So the polls were right in 46 to 48 out of 50 states. Two out of the three states Trump flipped - Michigan and Pennsylvania - were very close anyway and Trump won them by narrow margins - 13,000 votes (if that holds) in Michigan and 68,000 in Pennsylvania.


The big surprise was Wisconsin, where Clinton was supposed to win by 6 points. Trump ended up beating her by 27,000 votes. Okay, maybe the Wisconsin pollsters should feel embarrassed.

The fact of the Trump victory is of course massive. But do not misunderstand what happened. There was no landslide. No sea change. The demographic realities still make it very difficult for a Republican to win a national election. This was the sixth of the last seven elections where the Republicans lost the popular vote.

A Trump re-election effort in 2020 will probably be even more of a challenge.

Trump had to run the tables. He did it, of course. But for all the credit Trump and his team deserve for putting him in a position where he even had an opportunity to run the tables, to me, the win has more of the character of a lucky break.

Or, if you prefer, a miracle.